RIKEN finds that pleasant and unpleasant odor sensations are generated by different neural circuit structures in the lateral horn in the brain

A research team led by Special Postdoctoral Researcher Makoto Someya, Technical Staff I Kazumi Ohta, and Team Director Hokto Kazama at the RIKEN Center for Brain Science announced the discovery of brain cells that compute innate odor values (pleasant and unpleasant) and found that information about pleasant and unpleasant sensations is generated through different circuit mechanisms. By combining calcium imaging and various analytical methods using fruit flies, the team revealed that innate odor values are computed in the lateral horn of the brain. This discovery is expected to contribute to the realization of a digital twin of the brain. The results were published in Cell on September 17.

Animals innately act based on the value of sensory stimuli for survival, approaching pleasant odors and avoiding unpleasant ones. In the olfactory circuits of mammalian brains, the physicochemical properties of odors are encoded in the primary olfactory center, and it had been suggested that odor value information is extracted in the higher-order centers that receive this information.

However, cells encoding odor values in the higher-order centers had not been identified, and the details remained unclear.

The structure and function of the olfactory system are known to be highly conserved between mammals and insects. Fruit flies exhibit clear avoidance and attraction behaviors to odors, allowing evaluation of innate odor values for the organism. Genetic techniques are well established, making evaluation and manipulation possible, and the brain is small enough to comprehensively record the activity of cell populations. However, no method existed for comprehensive and specific recording of lateral horn neurons.

Therefore, in this study, the team developed a measurement technology combining calcium imaging and optogenetics using fruit flies.

As a result of measuring cell populations in the lateral horn and mushroom body, higher-order brain regions that process odor information, they found that two distinct neuronal populations in the lateral horn each encode specific odor values. One responds only to pleasant odors, and the other responds only to unpleasant odors. The magnitude of response correlated with the degree of pleasantness or unpleasantness.

When they created a mathematical model to predict odor values from the activity of neuronal populations, using lateral horn activity resulted in higher prediction accuracy than using the mushroom body, demonstrating that the lateral horn is the center for computing innate odor values.

To investigate this neural circuit mechanism, the team performed analysis using the connectome, a map of neural circuits representing detailed connections between elements within the animal's nervous system.

As a result, they found that neurons in the primary olfactory center that contribute to behavioral selection for pleasant and unpleasant odors were selectively connected to specific lateral horn cells. In the lateral horn cell population receiving inputs related to pleasant odors, inhibitory connections between lateral horn cells were prominently observed, suggesting that "pleasant" information is generated only after being further processed by local inhibitory circuits in addition to inputs from the primary olfactory center.

To verify this, they constructed and tested a neural network model based on the connectome. When actual odor responses from the primary olfactory center were input into the model, the activity of lateral horn cells that respond specifically to pleasant and unpleasant odors was reproduced. When local inhibition was removed, lateral horn cells responding to unpleasant odors remained, but those responding to pleasant odors disappeared, revealing that pleasant and unpleasant information is generated by different neural circuit structures.

Kazama commented: "It was an unexpected result that pleasant and unpleasant odors are processed in circuits with different connectivity patterns. In other words, we found that at the circuit level, 'like' is not simply the opposite of 'dislike.' Going forward, we want to continue elucidating the operating principles of brain circuits through comprehensive neural activity measurements and simulations of neural activity utilizing whole-brain wiring diagrams."